Any ideas of the cause? BIOS itself works fine in other systems, so it's a failure in the MVS hardware. Just not sure where it is...
If your sure the EPROM is not failing then its either going to be the 68k or something interfering with the data lines. Do you have known working game carts. If so get to the UniBios main menu, hilight the game cart check then hold start and press the A button. This will test the bios again with a stronger check. See if it always passes or sometimes fails. Then do the same check with a game cart and see if that sometimes passes and then other times fails.
Well that's just it, all carts work. I see this error about 75% of the time, but if I just hit <A> to resume, it carries on and everything works. Once I bypass the error, there is no sign that there's any problem. Unless I allow the hardware test to complete, then it's bricked. So to avoid that I have the battery pulled and UNIBIOS defaults to not running the tests.
He's not telling you to test the carts, the test he's telling you to do is to test the bios itself which uses the CRC check that is normally used for checking the game carts.
You have not read properly my post properly and your also going off the premise that if the game looks to run fine then it runs fine. That is not a good way to check that faults of this nature. Read my post again and do the checks I suggest in the way I suggest using the gamecartcheck menu. Well unless your happy with the fault as it is now and simply skippig it every time it happens.
Check if the bios on the motherboard has 38,39 and 40 connected to +5v, if not, connect them. That sometimes helps.
Yup, seems to have worked. However it found no problems. Hurmph, guess I have to break out the logic probe and check traces.
With smkdan having disappeared 5+ years ago, I am curious what folks would think about me releasing a disassembled version of the diag bios (sp1 + m1). I would like to make it available but it feels like a bit of a gray area since its not mine.
I did most of the disassembly a number of months ago as a way to learn 68k/z80 assembly, neo geo hardware and get to a better understanding of what some of the errors mean. Everything is fully disassembled, documented and compiles back to the same crcs as smkdan's released sp1/m1 files.
Last time I spoke to Dan he told me his plans for the Neo Diagnostics was to make a cart that would test the full system even full comm between different areas of the board. I'm not sure what happened to him as I've reached out multiple times, I hope nothing tragic as I was really looking forward to seeing it come to fruition.
Yeah, hes using 4 different tiles from sfix that are a solid color. Then using palettes to make them correct color/intensity when generating the color bars.
I will try and put up the disassembly tomorrow.